A thorough understanding of visual information processing is important not only because vision is the largest source of sensory input to the nervous system, but also because the visual system is a model for neural information processing in general. The proposed research continues our attempts to advance the understanding of complex visual processing and how the submodalities of vision interact through the use of visual evoked potentials and psychophysics. In project [1], the neural computations underlying the extraction of visual form will be studied. The working model for this process consists of two levels of nonlinear processing: one, a local subunit which rectifies a high-pass transformation of the visual input, and a second longer-range nonlinearity which combines the output of these subunits in a supra-additive fashion. It is hypothesized that the-second nonlinearity adjusts its behavior in a manner dependent on local contrast. A main focus of the proposed work will be on how this neural measure of contrast is generated and how it tunes the second nonlinearity. In project [2], the interaction of color and luminance signals will be studied. A particular focus will be whether the roles of color and luminance signals in form analysis are a direct consequence of the way in which chromatic and luminance pathways (viewed as linear filters) transform visual information, or whether more elaborate models are required. This analysis may also provide clues as to how chromatic signals present at retinal and geniculate levels are transformed into the opponent colors of psychophysics. In project [3], the neural computations underlying the extraction of motion will be studied. The investigation will be organized around the hypothesis that the model framework of [1], along with appropriate time-delays, accounts for the extraction of motion from standard stimuli as well as "non-Fourier" motion stimuli. All of these studies are based on novel and sophisticated visual stimuli and analytical techniques; further development of such techniques is implicit in the proposed research.